针对一种非常有效的信号识别算法——滑窗法(sliding window,SW),在其嵌入的串行信号识别流程基础上提出并行处理方案,并基于MPI(message passing interface)平台实现并行优化版本。这在基因芯片研究领域是个崭新的思路。最后,根据公共数据Affymetrix(2005) Tiling Array,展示SW并行流程的运行性能。实验结果表明,信号识别流程的并行化对大规模的数据运算非常有效且十分必要。
This paper focused on an effective signal identification algorithm,sliding window(SW),promoted parallel method on the basis of serial realization in which it was embedded and completed an optimal parallel program using MPI(message passing interface),which was a brand-new thought in the gene chip field.A conclusion was drawn from the result testing on the public Affymetrix Tiling Array data that parallel processing is necessary and has good effect on signal identification pipeline.